Playful Search: 1. Understanding the Query
Google aims to prioritize results that:
Address the correct or more likely intents of the query
Provide a satisfactory amount of content for the addressed intents
Do not have certain quality or relevance issues
In this article, I will explore the first aspect of search results relevance: searcher intent.
Introduction
Since the relevance of a search result is always connected to a search query, the first step in determining that relevance is to understand the query. That's what Google tries to accomplish in the first place.
Everything starts with a need from the searcher that must be fulfilled. I will call this either user intent, search/searcher intent, or query intent. These terms can be used interchangeably.
Google doesn’t make it clear how exactly it classifies intent. What is clear, however, is that the traditional classification used by more than 90% of SEOs when referring to search intents fails its purpose of allowing for good content decisions.
To simply divide intents in navigational, transactional and informational is not good enough, and I find it much more useful to classify them by clarity – very clear, somewhat clear, and not clear at all – and likelihood – correct, more likely, less likely, and very unlikely intents.
Very Clear Intent: The Correct Intent
Sometimes, it’s very easy to infer what the searcher intent is based on the issued query.
For example, <buy iphone 11 128gb>. What’s the user intent?
— To buy an iphone 11 with 128GB of storage capacity.
<iphone 11 models compared>. What’s the user intent?
— To compare between different models of the iPhone 11.
What about this one: <should I wait for phone 12>?
— It is clear that the query is misspelled (phone instead of iPhone), and the user intent is to know whether waiting for the release of the iPhone 12 model is a good idea.
So, in these examples, the searcher intent is very clear.
Whenever the searcher intent is very clear, I like to call it the correct intent.
Somewhat Clear Intent: More Likely, Less Likely, and Very Unlikely Intents
Now, what do you think of this query:
<iphone>
What is the user intent?
Well, this is not so clear like in the previous examples. The user might be looking for iPhones to buy (product results), nearby stores that sell iPhone (local business results), iPhone’s official page on Apple’s website, images of iPhones, etc.
In such cases, the searcher intent is just somewhat clear. Although it’s hard to tell what exactly they are looking for, it’s still possible to figure out several different results that could be relevant or useful to them.
Somewhat clear queries have many possible intents associated with them. Some of these intents are more likely, some are less likely, and some are very unlikely.
For example, for the query <iPhone>,
Some of the more likely intents are:
To look for an iPhone to buy online
To look for iPhone’s official website
To look for news about the iPhone
To look for nearby stores that sell the iPhone
Less likely intents:
To look for informational articles about the iPhone (e.g. history, models, reviews, etc.)
To look for images of the iPhone
Etc.
And very unlikely intents:
To learn how to fix a specific bug on iPhone
Etc.
How Google assesses the likelihood of possible searcher intents
The truth is that Google doesn’t even make it clear how intents are actually classified, let alone how it determines which intents are more likely than others. The answer lies in the bowels of its algorithms and, of course, in the SERP.
As Search theorists and researchers, we can presume that there are several different signals that play a role together for this purpose.
For example, if the search engine notices that many people who search for <iphone> end up clicking on news results, then it assumes that the intent to look for news is probably among the most likely ones for this query. The intent to look for news, not informational, transactional or navigational.
See how this simple, straight-to-the-point classification makes much more sense from a strategic perspective than the generic informational category. By calling the intent by what it means — in this case, to look for news — it is possible to evaluate the likelihood of this specific type of result to rank.
This classification also makes possible to assess the seasonality of such intents, among other things. For example, the news intent will probably become even more likely for the <iphone> query weeks before the launch of a new model — and Google will certainly boost news results when it figures this out.
When the search engine is not confident about the likelihood of intents for a specific keyword, there will likely be high volatility in the SERP, in an attempt to experiment and assess what kind of results seem more relevant to users.
It is also almost certain that algorithms infer the intent likelihood of certain queries based on other related queries for which Google has more user data. For example, if most people who search for the name of bands or singers end up clicking on video clips, then the intent to look for video clips is among the most likely ones for queries formed by just the name of a band or singer.
In the Search Algorithm section about How Search Works, Google says that "understanding intent is fundamentally about understanding language”, and that it tries to understand the type of query you’ve entered by applying some of the latest research on natural language understanding.
The importance of diversity
If Google aims to prioritize results that address the correct or more likely intents of the query, are results addressing less likely intents always bound to never make it to the first page of the search engine?
Not really, as the SERP usually becomes more relevant when diversity is to some extent represented on the page. This diversity can be in the form of content type, length, structure, and also intents that it addresses.
Take this query as an example: <seo>.
It seems reasonable to assume that the single most likely intent of this broad query is “to look for information about SEO”. For example, what it is, how it works, the different categories that make it up, etc.
As a user, would you prefer that the first page of Google displays a list of 10 different articles about what SEO is, how it works, etc., all bringing a similar value in terms of content, or would you rather be presented with more options, even if these results represent less likely intents of the query? For example, a map result that lists SEO agencies nearby, an article with the best SEO tools, etc.
In this case, it seems that a more diverse set of results would make the SERP richer for this query, and Google generally thinks so too. Therefore, it’s not uncommon to see results representing less likely intents ranking on the first page as well.
Typically, the search engine will look for diversity among results addressing most likely interpretations and intents. If the diversity from these results is not enough to generate true value, then it will likely diversify the SERP by also including less-likely-intent results in the first page.
